270 Prof. Chapman on the Classification of the Silicates 



Cubic expan- Determined 



Substance. Formula. sion for 1°. by means of 



n , , , , fKO, SiO 3 + Al 2 O 3 , f 0-000026 Water 



urtnojtias . . . ^ 3gi03 \0'000017 Mercury 



Glass, soft soda glass 0-000026 Water . 



Glass, soft soda glass, another kind . . 0-000024 Mercury 



Glass, hard potash glass 0-000021 Mercury 



Taking every possibility of error into account, M. Kopp con- 

 siders that we may infer with certainty from the preceding num- 

 bers, that the expansion of solid substances is by no means 

 determined by their chemical nature. The difference between 

 the coefficients of expansion for arragonite and calcareous spar is 

 so great as to destroy all hope of establishing any relation of the 

 kind. Neither does the expansion appear to depend altogether 

 on the arrangement of the atoms ; for although bitter spar and 

 carbonate of iron agree, and heavy spar differs but little from 

 ccelestine, in the cases of carbonate of iron and carbonate of lime, 

 and of rutile and oxide of tin, no such agreement exists. The 

 table further shows that there are many non-metallic substances 

 which expand as much under the action of heat as the metals 

 themselves. 



XXXVIII. On the Classification of the Silicates and their allied 

 Compounds. By Edwaro J. Chapman, Professor of Mine- 

 ralogy in University College, London*. 



IN the present state of chemical and mineralogical knowledge, 

 two distinct classifications of minerals seem to be necessary ; 

 the one, strictly chemical, having regard solely to the actual 

 composition of the substance, and being independent, conse- 

 quently, of all deductions based upon isomorphism or other mo- 

 difying causes ; and the other, a chemico-physical distribution, 

 founded upon a careful study and interpretation of the entire 

 nature of the mineral in all its bearings. The first shows us 

 what a mineral really is in its relations to existing chemistry, but 

 only in these relations ; being unable, on the one hand, amongst 

 other apparent contradictions, to define and separate isomero- 

 heteromorphous bodies ; and forced, on the other hand, to forego 

 natural analogies, in the separation of substances evidently akin 

 to one another. A classification of this kind is, nevertheless, 

 necessary : first, as a classification of convenience in a chemical 

 and ceconomical point of view ; and secondly, as a check upon 

 the too hasty generalizations to which the chemico-physical 

 system naturally gives rise. 



In the following classification I have attempted to develope a 

 chemico-physical distribution of the silicates and their allied 



* Communicated bv the Author. 



